Foraging distributions of little auks Alle alle across the Greenland Sea: implications of present and future Arctic climate change

(1)

INTRODUCTION

In recent years, the effects of climate change have been especially pronounced in the Atlantic sector of the Arctic. In the Greenland Sea, sea ice has declined and the eastern side has experienced large increases

in temperature (Blindheim et al. 2000, Vinje 2001).

These changes have been driven primarily by increased advection of warm, saline, Atlantic water into the Greenland Sea via the West Spitsbergen Current (WSC) (Aagaard et al. 1987, Walczowski & Piechura 2006, Fig. 1).

*Email: [email protected]

Foraging distributions of little auks Alle alle across the Greenland Sea: implications of present and

future Arctic climate change

Nina Karnovsky

^1,

*, Ann Harding

²

, Wojciech Walkusz

³

, S8awomir Kwas´niewski

³

, Ilona Goszczko

³

, Josef Wiktor Jr

³

, Heli Routti

⁴

, Allison Bailey

¹

, Laurel McFadden

¹

,

Zachary Brown

¹

, Grégory Beaugrand

⁵

, David Grémillet

^{6, 7}

1Department of Biology, Pomona College, 175 W 6th St., Claremont, California 91711, USA

2Environmental Science Department, Alaska Pacific University, 4101 University Drive, Anchorage, Alaska 99508, USA

3Institute of Oceanology, Polish Academy of Sciences, Powstan´co´w Warszawy 55, 81-712 Sopot, Poland

4Norwegian Polar Institute, Polar Environmental Centre, 9296 Tromsø, Norway

5Centre National de la Recherche Scientifique, Laboratoire d’Océanologie et de Géosciences, UMR CNRS 8187 LOG, Université des Sciences et Technologies de Lille 1, Station Marine, 28 avenue Foch, 69360 Wimereux, France

6Centre National de la Recherche Scientifique, CEFE, 1919 route de Mende, F 34293 Montpellier, Cedex 5, France

7Percy FitzPatrick Institute, DST/NRF Centre of Excellence, University of Cape Town, Rondebosch 7701, South Africa

ABSTRACT: The Arctic is undergoing widespread warming. In order to understand the impact of climate change on Arctic marine food webs, we studied the at-sea distribution of foraging little auks in contrasting conditions of the Greenland Sea. While the eastern side of the Greenland Sea has experienced recent warming, the western side is still dominated by cold, Arctic water in the East Green- land Current. We hypothesized that foraging little auks would be found in greatest abundance in cold Arctic waters bearing more lipid-rich prey, allowing them to deliver more energy-rich food to their chicks. To test our hypotheses, we made ship-borne bird observations and zooplankton tows, as well as analyses of chick meals at 2 little auk colonies adjacent to 3 distinct water masses in the Greenland Sea. Associated with the coldest water in the East Greenland Current, we found the highest concentrations of large Calanuscopepods (C. glacialisand C. hyperboreus), as well as the highest concentrations of foraging little auks, indicating a relationship that is likely to be disrupted by increasing water temperatures. To assess potential future impacts of ocean warming, we used a coupled atmosphere–ocean global climate model (AOGCM) to predict Greenland Sea sea-surface temperatures over the study area at the end of the 21st century. Our results suggest that 4 of 8 little auk breeding colonies in the North Atlantic may be negatively impacted as temperatures exceed the thermal pref- erenda of large Calanus, which is the preferred prey of little auks during the breeding season.

KEY WORDS: At-sea observations · Calanuscopepod · Climate model · East Greenland Current · Seabird · Little auk · Sea temperature · West Spitsbergen Current

Resale or republication not permitted without written consent of the publisher

O PEN PEN

A CCESS CCESS

(2)

The Greenland Sea is made up of diverse water masses. Briefly, the WSC originates in the Atlantic Ocean and flows north alongside the cold Sørkapp Current (SC; Saloranta & Svendsen 2001). The SC also runs north along the west coast of Spitsbergen inshore of the WSC but advects cold, low salinity, Arc- tic water that originates northeast of the island (Swer- pel & Zajaczkowski 1990; Fig. 1). The East Greenland current (EGC) is characterized by cold, low salinity water. It originates in the Arctic Ocean and flows south along the east coast of Greenland (Bourke et al.

1988; Fig. 1).

A different numerically dominant Calanuszooplank- ton species can be found in each of these water masses.

The EGC primarily supports the large, lipid-rich copepod Calanus hyperboreus (Hirche 1997). The SC is dominated by medium-sized C. glacialis(Karnovsky et al. 2003). The WSC advects high densities of C. fin- marchicus, which is the smallest, least lipid-rich spe- cies of the Calanuscopepods, into the Greenland Sea (Scott et al. 2000, Karnovsky et al. 2003). Some C. fin- marchicusare recirculated in the Greenland Sea and the Arctic Ocean and get mixed into both the SC and the EGC (Hirche & Kosobokova 2007).

Such marked differences shape regional food webs, and the distribution and ecology of planktivorous organisms, including marine mammals and seabirds.

For instance, Karnovsky et al. (2003) found that little auks Alle alle, which are small planktivorous seabirds, responded strongly to the distribution of water masses and associated copepods on the east side of the Green- land Sea. Little auks breeding in Hornsund Fjord on

the southwest side of Spitsbergen foraged in the SC where they fed on energy-rich Calanus glacialis and avoided the WSC where the smaller C. finmarchicus was found in high densities (Karnovsky et al. 2003). In a similar fashion, Hunt & Harrison (1990) found that least auklets Aethia pusillaon King Island, Bering Sea, flew to feed in a water mass that had larger, more energy-rich prey.

Recent warming of the eastern sector of the Green- land Sea has been particularly pronounced in both the SC and the WSC. During the period 1965 to 1997, the SC has warmed by 1°C (Blindheim et al. 2000) and surface water temperature of the WSC increased by as much as 2°C (Hurrell 1995, Schlichtholz & Goszczko 2006). In contrast, the cold, low salinity water flowing from the Arctic Ocean appears to have been a rela- tively constant feature in the western sector (as sum- marized by Hop et al. 2006).

Little auks are the most abundant seabird species in the North Atlantic (80 million ind., estimated from Kampp et al. 1987, Isaksen & Gavrilo 2000, Egevang et al. 2003) and are an essential component of local marine food webs. They occur on both sides of the Greenland Sea (Spitsbergen and East Greenland) where the 2 populations experience the highly contrasting environments of the EGC and the SC/WSC.

Comparison of these 2 populations allowed us to test the impact of warming in the Greenland Sea on their foraging conditions.

More specifically, our objectives were to (1) compare the at-sea distribution and abundance of little auks foraging in both the eastern and western sectors of the Greenland Sea, (2) assess the physical properties of the water masses where the birds did, and did not, forage, (3) determine the species composition, size and density of prey in the different water masses, (4) assess the impact of foraging in different water masses on chicks at colonies on either side of the Greenland Sea in terms of the energy they receive, and (5) model future temperature increases in the region and assess how it may impact the distribution of zooplankton species, with energetic implications for little auks.

Given the physical and biological properties of the different water masses, we hypothesized that there would be strong zonal differences in (1) the foraging distribution of little auks, (2) the composition of diets fed to chicks, and (3) the amount of energy delivered by little auk parents to their chicks. We predicted that birds foraging on the western side of the Greenland Sea would forage in the EGC on the largest calanoid copepod Calanus hyperboreus. On the eastern side, we predicted that little auks would forage in the SC and consume primarily the medium sized C. glacialis as in 2001 (Karnovsky et al. 2003). We predicted that the chicks being fed on the western side of the Green- Fig. 1. Study area. Arctic-fed East Greenland Current (EGC,

white arrow), Atlantic-fed West Spitsbergen Current (WSC, black arrow), and Arctic-fed Sørkapp Current (SC, grey

arrow)

(3)

land Sea would receive more energy than chicks on the eastern side. With observations and modeling results, we sought to answer the question: Will future increases in warming in this region result in a north- ward progression in the ranges of copepod species, resulting in a lower-energy zooplankton community dominating waters adjacent to several massive little auk colonies?

To test our hypotheses, we studied little auks breeding at and foraging off 2 colonies, one on each side of the Greenland Sea, in summer 2005. The Kap Höegh little auk colony is located on the east coast of Green- land near Scoresbysund Fjord (78° 43’ N, 22° 38’ W), which is an area influenced by the EGC (Bourke et al.

1988; Fig. 1). The Hornsund Fjord colony is located on the west coast of Spitsbergen (77° 15’ N, 15° 22’ E), which is adjacent to the WSC and the SC (Aagaard et al. 1987, Karnovsky et al. 2003; Fig. 1).

Little auks are planktivorous seabirds that migrate to the Greenland Sea during their breeding season to feed in its productive waters. Throughout their range, during the breeding season little auks mainly feed their chicks with calanoid copepods of the genus Calanus(e.g. Roby et al. 1981, Weslawski et al. 1999a, Pedersen & Falk 2001, Karnovsky & Hunt 2002, Karnovsky et al. 2003, Karnovsky et al. 2008). Little auks are subsurface foragers that use their wings for underwater propulsion, diving to maximum depths of 35 m (Falk et al. 2000). They raise one chick during the breeding season and both males and females partake in feeding their chick. A pair of little auks feeds its chick ~5× d^–1 (Stempniewicz 2001) and each food delivery can contain over 1000 prey items (Karnovsky et al. 2003). Adult little auks collect prey for their chick in a specialized gular pouch (Stempniewicz 2001), and chicks leave the colony when they reach 25 d of age on average (Harding et al. 2004).

The 3 aforementioned Calanusspecies (C. finmarchi- cus, C. glacialis, and C. hyperboreus)present strong contrasts in size and energy content. Previous studies of gular pouch content have determined the size/stage range commonly taken by little auks: C. finmarchicus copepodid stage CV and adult females (AF); C.

glacialiscopepodid stage CIV, CV, and AF; C. hyper- boreus CIV, CV, and AF (hereafter ‘edible’ Calanus;

We˛s8awski et al. 1999a,b, Karnovsky et al. 2003). Scott et al. (2000) found that stage V of C. hyperboreusand C. glacialiscopepodids contain 25× and 8× the lipid content of stage V C. finmarchicus, respectively.

Making this comparative study of the contemporary Greenland Sea, and modeling future copepod distributions based on predicted ocean temperature changes allowed us to gain insight into how present and future warming may impact Arctic marine food webs in the Greenland Sea.

MATERIALS AND METHODS

At-sea surveys.We conducted at-sea surveys adjacent to both Kap Höegh, eastern Greenland (hereafter:

Greenland) and Hornsund Fjord in southwestern Spitsbergen (hereafter: Hornsund) in 2005. We conducted surveys when chicks at both colonies were

~15 d old during the mid-chick rearing period. We conducted the Greenland at-sea surveys aboard the sailboat RV ‘Vagabond’ (15.3 m) from 1 to 11 August, and the Hornsund surveys from 29 to 30 July aboard the sailboat RV ‘Oceania’ (48.9 m). Despite the slight differences in timing, both cruises were conducted during the same stage of chick rearing at each colony. We conducted surveys whenever the boats were under- way and counted foraging little auks found in a 90° arc from directly in front of the boat out to one side (to 300 m) off the side with the best visibility (following the methods described by Karnovsky et al. 2003). Our survey track on the eastern side of the Greenland Sea was designed to cross the shelf break at 500 m depth and to repeat the surveys conducted in 2001 (Karnovsky et al.

2003). On the western side of the Greenland Sea, we designed our survey lines to make several zonal tran- sects across the EGC. We counted all little auks that were sitting on the water because we considered these birds to be foraging instead of transiting (flying) to another foraging site. We recorded our observations directly into a computer. We calculated densities of foraging birds (number of sitting or feeding birds km^–2) along 3 km segments of the transect lines. We believe that the little auks seen during the at-sea surveys orig- inated from the colonies adjacent to the survey areas because of the very short foraging trips that birds make to collect food for their chicks (Welcker et al.

2009b), and because in 2001, Karnovsky et al. (2003) found little auks foraging within 50 km of the Horn- sund colony.

Zooplankton and physics sampling.To assess prey species and densities, we made vertical plankton net tows from 50 m depth using a WP-2 net with a 0.25 m² opening area (Tranter & Fraser 1968) and fitted with filtering gauze of 500 µm mesh size. The nets were fitted with a mechanical flow meter (General Oceanics, model 2030). The only difference in the sampling methods between the 2 sites was that a hand winch was used at the Greenland site whereas a mechanical winch was used at the Hornsund site. We preserved the zooplankton in 5% formalin and seawater buffered with borax and analyzed the samples under a dissect- ing microscope. We identified plankton to the species level and life stage whenever possible and measured taxa that varied in length. We calculated the densities (m^{– 3}) of the edible Calanus species and stages. We used reported values of dry weights for Calanusspe-

(4)

cies and life stages (Karnovsky et al. 2003 and sources therein) and an energetic value estimate of 26 kJ g dry wt^–1 (We¸8sawski et al. 1994) to calculate the energy available to little auks in the different water masses.

To determine the physical characteristics of the water masses in the study area, we stopped surveys every 10 to 15 km to conduct conductivity- temperature-depth (CTD) casts (see locations of circles in Fig. 3). We used the water temperature and salinity obtained at 15 m depth in all analyses. We classified water masses as EG, WSC, or SC based on temperature, salinity, and location.

Chick diets and provisioning frequency.In order to compare the composition of prey taken by little auks at the 2 colonies, we used mist nets and noose carpets to catch adult little auks as they were returning to the colony with food in their gular pouch. Once caught, we gently scooped out the prey from their gular pouch using a small plastic spoon and preserved the meals in 5% formalin and seawater. We banded, measured and weighed the adults before they were released, and no bird was caught more than once. We counted and identified all chick diet items to the lowest possible taxon and estimated total energetic content, again using reported values of dry weights of the various taxa (Berestovskii et al. 1989, Mumm 1991, Hanssen 1997, We¸8sawski et al. 1999a, Karnovsky et al. 2003) and an energetic value estimate of 26 kJ g dry wt^–1 (We¸8sawski et al. 1994). We only included complete diet samples in the analyses of the energetic content, percent composition and average number of prey items per food load. We calculated the percent composition by dividing the number of individuals of the taxon of interest by the total number of zooplankters collected for each colony.

We assessed the number of food loads chicks receive per day at both colonies by observing marked provisioning birds for 24 h during the mid-chick rearing period when chicks were ~10 to 15 d old (as in Welcker et al. 2009b). These observations were made during (Greenland) or just prior to (Hornsund) the oceanographic sampling. We caught provisioning birds in their nests or just outside of their nests with noose carpets. Each bird was marked on its breast feathers with a unique colour pattern and given a unique combina- tion of coloured rings on their legs. We sat ~10 to 15 m from the nesting area for 24 h and recorded the frequency and timing of entry of the marked little auks into their nests with food for their chicks. A minimum of 2 observers collected these data simultaneously to ensure that all feeding events were recorded.

Spatial analysis and habitat associations. We com- pared the spatial distributions of ocean temperature, zooplankton abundance, and density of foraging little auks using ArcGIS. Bins of bird density were overlaid

over a temperature raster (or krige) for the study area based on the temperatures measured at 15 m from the CTD casts. We repeated this process with zooplankton densities overlaid over the temperature raster.

Modeling the impact of future climate change.We used data from the continuous plankton recorder (CPR) to calculate the thermal preference of the 3 important Calanusspecies in the Nordic seas: C. fin- marchicus, C. glacialisand C. hyperboreus(Warner &

Hays 1994, Beaugrand et al. 2007, Helaouët & Beau- grand 2007). The mean abundance per CPR sample of each of these copepod species was calculated for each temperature category (–2 to 25°C by 1°; see Fig. 7) using the International Comprehensive Ocean–

Atmosphere Data Set (ICOADS 1 Degree) provided by NOAA/OAR/ESRL PSD, Boulder, Colorado, USA (Woodruff et al. 1987).

To assess the potential impact of changes in sea surface temperature (SST) and water masses on little auk colonies by the end of the 21st century, data from the European Centre at Hamburg (ECHAM 4) were used in a coupled atmosphere–ocean general circulation model (AOGCM; Roeckner et al. 1996) with a horizon- tal resolution of 2.8° latitude and 2.8° longitude. Pro- jections of monthly SST from 1990 to 2100 were based on scenario A2 (atmospheric pCO2 of 856 ppmv by 2100; Intergovernmental Panel on Climate Change Working Group Report I 2001; see Beaugrand et al.

2008 for information on the nature of these data). In addition, we used SST and ice data (HadlSST 1.1) from 1990 to 2005 to examine the mean location of sea ice in relation to little auk colonies. These data had a spatial resolution of 1° longitude and 1° latitude and were based on in situobservations of SST and sea ice as well as satellite derived estimates (AVHRR for SST and SSMI for ice) as in Rayner et al. (2003).

RESULTS

We performed CTD casts and net tows at 25 stations adjacent to each colony. In Hornsund, these represented 9 WSC and 16 SC stations. In all, 9510 little auks were observed foraging in the waters off Horn- sund in 128 three km transect segments, and 8584 in the waters off Greenland in 236 three km segments.

Water temperatures, zooplankton and foraging little auks

The distribution of ocean temperatures off Hornsund indicated the presence of 2 ocean currents with distinct temperatures (Fig. 2). Inshore was a mass of cold water that can be identified as the Arctic-fed SC, with tem-

(5)

peratures ranging from 2.4°C inshore to 5.5°C along the front with the WSC. Offshore was the Atlantic-fed WSC, with a temperature range of 5.5 to 9°C (Fig. 2).

There were significantly higher densities of Calanus

finmarchicusin the WSC than in the SC (1-tailed t-test p = 0.003) (Fig. 3, Table 1). Conversely, there were significantly higher densities of large C. glacialis in SC than in WSC water (1-tailed t-test p = 0.013) (Fig. 3,

Scoresbysund Fjord

Hornsund Fjord

EGC

SC

WSC 20°W

0.12 8.99

Temperature (°C)

Little auk density (ind. km^–2) –1.30 2.41

0 0.1–10 10.1–25 25.1–50 50.1–100 100.1–250 250.1–500 500.1–850

14°E 15°E

77°N

71°N

10 km 10 km

22°W

300 m

Fig. 2. Alle alle. Densities and distribution of foraging little auks in relation to water mass temperatures (measured at 15 m depth) adjacent to study colonies. Abbreviations are as in Fig. 1

0 0.1–10 10.1–25 25.1–50 50.1–100 100.1–250 250.1–500 500.1–850

20 km

WSC SC

EGC

300 m 300 m 300 m

Calanus finmarchicus

Temp. (°C)Temp (°C)

Density of zooplankton

(ind. m^–3)

Calanus glacialis Calanus hyperboreus

8.99

2.41

0.12

–1.30

Hornsund

Greenland

Fig. 3. Calanus finmarchicus, C. glacialisand C. hyperboreus. Densities and distribution in relation to water mass temperatures (measured at 15 m depth) adjacent to study colonies. Abbreviations are as in Fig. 1

(6)

Table 1). C. hyperboreuswas present in low densities (Fig. 3, Table 1), and was also more closely associated with SC than with WSC water (1-tailed t-test p = 0.028). The highest densities of foraging little auks (max 778 ind. km^–2) were found associated with SC waters, and hence with high densities of C. glacialis (Figs. 2 & 3). Little auks were not found foraging where there were high densities of C. finmarchicus.

Unlike Hornsund’s 2 distinct Arctic- and Atlantic-fed currents, the water temperatures off Greenland indicated a single very cold water mass, with a temperature range of –1.30 to 0.12°C (Fig. 2). The densities of Calanus species were more evenly distributed throughout the Greenland study area, and the densities of foraging little auks were not found to be associated with the presence of any one Calanusspecies. In general, zooplankton species and foraging little auks were evenly distributed throughout the sampling area, but with exceptionally high little auk densities occurring near the mouths of east Greenland fjords (Fig. 2).

The waters off Greenland contained far higher numbers of edible C. hyperboreus than did either water mass near Hornsund (Table 1, Fig. 3).

The major energy source available to little auks in the SC was Calanus glacialis, whereas in the WSC, the majority of energy was in the form of C. finmarchicus (Table 1). In the EGC, both C. glacialisand C. hyper- boreuscontribute a large portion to the energy content of the water (Table 1). Overall, the total energy provided by edible (large size) life stages of Calanus species was highest in the WSC (mean ± SE, 1.38 ± 0.27 kJ m^{– 3}, due largely to the high density of C. finmarchicus; Table 1).

Chick diets

We collected 27 complete chick meal samples in Greenland from 25 July to 7 August, and 33 in Horn- sund (22 to 28 July & 1 August). Diet samples from the 2 sites contained similar numbers of food items (Green- land: 920 ± 88; Hornsund: 1010 ± 101; p = 0.516). The compositions of diet samples at the 2 colonies were similar in that they were both dominated by copepods of the genus Calanus; however, the relative contribu- tions of C. hyperboreus, C. glacialis, and C. finmarchi- cus differed between the colonies. The chick diets at Greenland were dominated by C. hyperboreus (57.6%) and C. glacialis (32.3%) (Fig. 4), while food loads collected in Hornsund were overwhelmingly C.

glacialis (86.7%), with C. hyperboreus being nearly absent from the diet (Fig. 4). C. finmarchicusoccurred in low numbers in the diet, totaling 6.0% of the diet items in Greenland and 2.9% in Hornsund. Non- calanoid taxa represented 4.1% of total diet items in Greenland (of which Apherusa glacialiswas the dominant species) and 9.8% in Hornsund (of which Pagurus pubescenszoea was dominant) (Fig. 4). The average energy in a chick meal did not differ between Horn- sund and Greenland (Hornsund: 17.2 ± 1.4 kJ; Green- land 16.0 ± 1.2 kJ; 2-tailed t-test p = 0.531). We found no difference in the number of meals per day that little auk chicks in East Greenland received and that which chicks in Hornsund received (5 ± 1.1 and 3.5 ± 0.8 meals d^–1 for Greenland and Hornsund, respectively;

2-tailed t-test p = 0.13).

Modeling results

The CPR survey data were collected south of the study area; however, the surveys spanned Atlantic water masses over the entire ocean temperature range encountered. We found that Calanus finmarchicuswas preferentially located in waters that are 5 to 7°C, while the optimum temperature for C. glacialisand C. hyper- boreuswas much lower, at 0 to 2°C (Fig. 5). Neglecting areas that are ice-covered one or more months of the year, our modeled SSTs in the Nordic Seas for the period 1990 to 2005 matched well with observations (Fig. 6a,b). By extending this model into the late 21st century using current trends in levels of anthropogenic forcing (scenario A2), we project temperature increases across the entire little auk breeding range, with the greatest increases occurring in what are cur- rently the coldest surface waters of the study area (Fig. 6c,d). By breaking these modeled results into 3 temperature ranges (≤3°C for the C. glacialisand C.

hyperboreusthermal preferendum, ≥5°C for the C. fin- marchicus thermal preferendum, and 3°C to 5°C for Species Density in net tows Range Available energy

Water mass (ind. m^{– 3}) (ind. m^{– 3}) (kJ m^{– 3}) C. finmarchicus

WSC 179.8 ± 82.2 3.3–849.6 1.11 ± 0.46

SC 7.9 ± 2.8 0.0–39.6 0.04 ± 0.01

EGC 26.1 ± 6.1 3.0–128.3 0.15 ± 0.05 C. glacialis

WSC 36.1 ± 10.6 0.0–100.8 0.28 ± 0.08 SC 100.2 ± 17.8 15.6–258.3 0.78 ± 0.14 EGC 36.8 ± 7.0 1.8–162.0 0.40 ± 0.10 C. hyperboreus

WSC 0.5 ± 0.3 0.0–3.60 0.00 ± 0.00

SC 1.3 ± 0.5 0.0–7.60 0.02 ± 0.01

EGC 28.5 ± 6.4 3.0–169.9 0.41 ± 0.12 Table 1. Calanus finmarchicus, C. glacialis and C. hyper- boreus. Densities and energetics (means ± SEs) of edible large-stage copepods (C. finmarchicus: copepodid stage, CV, and adult females, AF); C. glacialis: stages CIV, CV, AF; C.

hyperboreus: stages CIV, CV, AF) in 3 divergent water masses (WSC: West Spitsbergen Current, no. of tows = 10; SC:

Sørkapp Current, no. of tows = 15; EGC: East Greenland Current, no. of tows = 25)

(7)

the biogeographical boundary), we tailored the SST model expressly for the effects it may have on little auk populations during the peak of the breeding season, i.e. July, when foraging demands are at their highest.

Toward the end of the 21st century, 4 of 8 major (>10 k breeding pairs) little auk colonies in the modeled area (Stempniewicz 2001) may face a shift from large/

energy-rich to small/lean copepods (Fig. 7).

DISCUSSION

We identified marked differences in the physical and biological properties of water masses between and within our 2 study areas on opposite sides of the Greenland Sea. Cold, low salinity Arctic waters of the EGC and the SC contained more large, energy-rich copepods than warmer, more saline waters of the WSC. Contrary to our hypothesis, the WSC had the highest total energy available from copepods due to high densities of Calanus finmarchicus. However, this species represented only a small portion of the diet of little auk chicks (2.9% of diet items in Hornsund).

Therefore, cold Arctic waters of the SC and the EGC are practically the most profitable waters to foraging little auks because they provide access to large, energy-rich copepods (Harding et al. 2009a and see below).

These oceanographic patterns were correlated with marked differences in the foraging distribution of little auks on each side of the Greenland Sea (Fig. 2). Off

Hornsund, birds largely restricted their foraging to the Arctic waters of the SC, confirming that their foraging preferences were associated with the abundance of Calanus glacialisrather than the total energy available from Calanus copepods in the water column. This pattern is consistent with results from at-sea surveys conducted in 2001 within the same zone (Karnovsky et al. 2003). In contrast, off East Greenland, little auks were found foraging throughout the study area, with high numbers at the mouths of east Greenland fjords (Fig. 2). This suggests the presence of mechanisms cre- ating areas of higher concentrations of prey, by physical forcing or due to increased biological production in hydrographic fronts, between the fresh glacial runoff and the more saline, denser seawater.

On both sides of the Greenland Sea, adult little auks provisioning chicks primarily consumed the larger copepod species available: Calanus glacialis on the eastern side (although C. finmarchicus was more abundant nearby) and C. hyperboreuson the western side (although C. glacialiswas equally abundant). On

50 40 30 20 10 0 0.8 0.6 0.4 0.2 0.0 0.6

0.4

0.2

0.0-2 0 2 4 6 8 10 12 Temperature (°C)

Calanus finmarchicus

Calanus glacialis

Calanus hyperboreus

Mean abundance per CPR sample

14 16 18 20 22 24 26

Fig. 5. Calanus finmarchicus, C. glacialisand C. hyperboreus.

Thermal associations of the 3 species found in continuous plankton recorder (CPR) samples

C. finmarchicus C. glacialis C. hyperboreus Other

Pagurus pubescens zoea Themisto sp.

Apherusa glacialis Other

Hornsund (n = 33)

Relative abundanceRelative abundance ‘other’

Greenland (n = 35)

87% 58%

72%

10% 18%

32%

6%

4%

58%

20%

21%

1%

3%

10%

0.6%

Fig. 4. Alle alle. Relative abundance (by number) of chick diet items. The upper pie charts represent all diet items, while the lower represent that fraction not consisting of Calanus

copepods

(8)

90°N 15 10 5 0 –5 –10 –15

15 10 5 0 –5 –10 –15

8 7 4 5 4 3

1 0 2 80°

70°

60°

50°100°W 50° 0° 50° 100°E

90°N

80°

70°

60°

50°100°W 50° 0° 50° 100°E

90°N

80°

70°

60°

50°100°W 50° 0° 50° 100°E

90°N

80°

70°

60°

50°100°W 50° 0° 50° 100°E

Sea Ice Modelled SST 1991–2005

a b Observed SST 1991–2005

Modelled SST 2080–2100

c d Difference a–c

Fig. 6. Sea surface temperatures (SST) for the period 1990–2005 (modeled and observed data) and projected changes in SST for the period 2080–2100 (scenario A2). (a) ECHAM 4 skin temperature (SST) for the period 1990–2005. Note that sea ice is not represented. (b) Observed SST for the period 1990–2005 using Hadley (HadlSST 1.1) data. The last 3 colours in the colour scale represent areas covered by sea ice, for the period 1999–2005, as follows: dark lilac: < 1 mo yr^–1on average, grey: between 1 and 4 mo yr^–1on average, pale lilac: ≥4 mo yr^–1on average. (c) ECHAM 4 skin temperature (SST) for the period 2080–2100. Note that sea ice is not represented. (d) Differences (°C) in modeled skin temperature (SST) between the period 2080–2100 (c) and the period 1990–2005 (a). Black circles: Large colonies of little auks (Stempniewicz 2001). Largest circle (Northeast Greenland):

33 million breeding pairs, medium circles: 1 to 10 million pairs, and smallest circles: 10 000 to 1 million pairs

90°N >7

5

3

<1

>7

5

3

<1 80°

70°

60°

50°

100°W 50° 0° 50° 100°E

90°N

80°

70°

60°

50°

100°W 50° 0° 50° 100°E

a b

Fig. 7. Location of 3 water masses between 1990–2005 (a) and 2080–2100 (b) using ECHAM 4 A2 data for July. Water masses in red have temperatures > 5°C and are therefore more likely to contain high proportions of Calanus finmarchicus(see Fig. 3). Water masses in blue have temperatures < 3°C and are therefore more likely to contain high proportions of C. glacialisand C. hyper- boreus. Water masses in yellow have temperatures ≥3°C and ≤5°C and represent the biogeographical boundaries. Grey circles:

locations of colonies, black circles: little auk colonies that will likely be negatively impacted by future temperature increase, green circles: colonies that will remain within the preferred temperature range of their preferred prey over the modeled time

frame. See Fig. 6 for the definitions of circle sizes

(9)

the eastern side of the Greenland Sea, the energy-rich WSC with high abundances of C. finmarchicus was well within the foraging range of little auks (only 50 km away from the colony). Brown (1976) reported finding little auks 100 to 150 km from their breeding colonies. Therefore, little auks clearly avoided areas where C. finmarchicuswas the dominant prey.

Finally, contrary to our predictions, and despite the aforementioned prey differences, the amount of energy delivered to little auk chicks per meal did not differ between the 2 study colonies. Furthermore, there were no significant intercolony differences in the average number of prey items per chick diet. One explanation for this result is that the little auks foraging adjacent to the Kap Höegh colony took more Calanus finmarchicus to their chicks than did Horn- sund little auks. In the EGC, C. finmarchicuswas as abundant as C. hyperboreus (Table 1). Therefore, while the energy content of most of the prey (C. hyper- boreus) was greater in the chick food loads at Kap Höegh, these birds also took higher numbers of the less energy-rich C. finmarchicus. We found no significant difference in the number of food deliveries between colonies; however, our observations were made for only 24 h on a small number of nests (Green- land 8, Hornsund 6).

Impacts of present oceanographic changes Our results show that little auks from the Greenland Sea experience diverse foraging conditions, but are ca- pable of providing similar amounts of energy to their chicks despite these contrasting abiotic and biotic conditions. The capacity to exert such plasticity has been demonstrated for a wide range of other seabird species confronted with environmental change (Golet et al.

2000, Hedd et al. 2002, Abraham & Sydeman 2004, Gall et al. 2006) and has also been demonstrated experimen- tally for little auks (Harding et al. 2009b). Nevertheless, there are indications that foraging conditions of little auks in the eastern part of the Greenland Sea are de- grading rapidly. Since 2001, little auks foraging on the eastern side of the Greenland Sea (adjacent to Spitsber- gen) have had variable diets (Jakubas et al. 2007, Mo- line et al. 2008), which could be associated with recent changes in the physical properties and distribution of the WSC and the SC and the zooplankton community these water masses support. Between 2001 and 2005, there has been a dramatic 7-fold decrease in the density of Calanus glacialisin the SC adjacent to Hornsund Fjord. In 2001, the density of C. glacialisin the SC was 716 ± 122 ind. m^{– 3} (Karnovsky et al. 2003). In 2005, there were only 100 ± 18 ind. m^{– 3}(n = 15; Table 1). Fac- tors driving this decline are unclear, but may include

the onshore flow of warm water from the WSC into the SC, or declines in ice cover and/or currents favorable to C. glacialis. Despite the decline in densities of C.

glacialis, this copepod was still the dominant species being carried back to chicks on the eastern side of the Greenland Sea. Hence, the threshold of minimal density of C. glacialisthat would force little auks to feed on C. finmarchicuswas not reached in 2005, but may well be reached in the future. Beyond a certain threshold of prey availability, Arctic waters may become unprof- itable for foraging little auks, with consequences for their reproductive success and population dynamics.

Although little auk parents have been shown to have flexibility in both their energy expenditure (Welcker et al. 2009a) and foraging behaviour (Welcker et al.

2009b, Karnovsky et al. unpubl.), life history theory predicts a limit to the additional reproductive costs breeders can accept, with parents prioritizing self- maintenance over increased provisioning effort when foraging costs become too high (e.g. Velando & Alonso- Alvarez 2003).

By comparing the zonal differences in water mass characteristics, zooplankton density, and distribution of foraging little auks, we gained insight into how increases in sea water temperature and consequently in the abundance of smaller prey may impact little auks. This spatial comparison allows us to understand how warming may impact little auks and other upper trophic predators that rely on Arctic copepods (e.g. the Arctic cod Boreogadus saida, which is a crucial link to Arctic upper trophic predators; Lønne & Gulliksen 1989).

Impact of future climate change

The distribution and strength of currents in the Greenland Sea and the interactions between them vary inter-annually (Loeng et al. 1997). They are also influenced by fluctuations of the North Atlantic Oscil- lation (NAO; Dickson et al. 2000). An increase in the influx of Atlantic water into the Arctic Ocean has been linked to a positive-phase NAO (Dickson et al. 2000).

Since the 1960s, the NAO has been in a predominantly positive phase with exceptionally high magnitudes (Hurrell 1995). Although the NAO is a natural mode of climate fluctuation, it is possible that it has been influenced by anthropogenic factors, such that the positive phase has become increasingly common (Visbeck et al.

2001). Increased Atlantic flow into the Arctic Ocean may reduce the extent of the SC near Hornsund Fjord, and thus the densities of the preferred prey, Calanus glacialis.

Based on our SST modeling results and the optimal temperatures of the prey species involved, it is clear

(10)

that many little auk colonies in the Nordic Seas will face a shift towards a zooplankton community dominated by Calanus finmarchicusby the end of the 21st century. We predict a greater temperature effect (impacting more little auk colonies) later in the season, which is the time of late chick-rearing and most inten- sive foraging by little auk parents and fledglings.

These findings mesh with predictions of Falk-Petersen et al. (2007) and Stempniewicz et al. (2007), who fore- cast that the eastern side of the Greenland Sea will become similar to the marine ecosystem off the west coast of Norway, with energy flow through C. fin- marchicus to herring Clupea harengus, rather than through C. glacialis to little auks. In the North Sea, common murres Uria aalgesuffered reproductive failure or had chicks fledge in poor body condition due to consumption of prey with low calorie content (Wanless et al. 2005). Climate change has forced a number of high latitude top predators to switch to sub-optimal prey, which has induced declines in adult body mass, chick fledging mass, adult survival, and reproductive success (as reviewed by Österblom et al. 2008). In a similar fashion, declines in Atlantic cod Gadus morhua L. in the North Sea are linked to shifts in the Calanus copepod community from C. finmarchicus to C. hel- golandicus associated with warming temperatures (Beaugrand et al. 2003). Further investigations are required to determine the exact threshold of levels of large copepod availability below which little auks will not be able to compensate via foraging plasticity, and will experience a reduction in chick and/or adult survival.

Acknowledgements. We thank the captains and crews of the RV ‘Oceania’ and RV ‘Vagabond’ (E. Brossier and F. P. du Sel), the 2005 station leader and expedition members at the Polish Polar Station at Hornsund Fjord, and Nanu Travel for support with logistics in East Greenland. All fieldwork was conducted with the permission of the Greenland Home Rule, Ministry of Environment and Nature, the Norwegian Animal Care Committee and the IACUC of Pomona College. Field- work was supported by the French Polar Institute Paul-Emile Victor (Grant #388 to A.H. and D.G.); the National Science Foundation (Grant # 0612504 to N.K. and A.H. and #0301469 to N.K.); Pomona College Summer Undergraduate Research Program (grant to A.B.); and the Mellon Foundation (grants to L.M. and Z.B.).

LITERATURE CITED

Aagaard K, Foldvik A, Hillman SR (1987) The West Spitsber- gen Current: disposition and water mass transformation.

J Geophys Res 92:3778–3784

Abraham CL, Sydeman WJ (2004) Ocean climate, euphausi- ids and auklet nesting: inter-annual trends and variation in phenology, diet and growth of a planktivorous seabird, Ptychoramphus aleuticus.Mar Ecol Prog Ser 274:235–250 Beaugrand G, Brander KM, Lindley JA, Souissi S, Reid PC

(2003) Plankton effect on cod recruitment in the North Sea. Nature 426:661–664

Beaugrand G, Lindley JA, Helaouët P, Bonnet D (2007) Macroecological study of Centropages typicus in the North Atlantic Ocean. Prog Oceanogr 72:259–273 Beaugrand G, Edwards M, Brander K, Luczak C, Ibañez F

(2008) Causes and projections of abrupt climate-driven ecosystem shifts in the North Atlantic. Ecol Lett 11:

1157–1168

Berestovskii EG, Anisimova NA, Denisenko CG, Luppova EN, Savinov VM, Timofeev CF (1989) Relationships between size and body mass of some invertebrates and fish of the North-East Atlantic. Academy of Sciences of the USSR. Murman Marine Biological Institute, Apatity Blindheim J, Borovkov V, Hansen B, Malmberg SAa, Turrell

WR, Osterhus S (2000) Upper layer cooling and freshening in the Norwegian Sea in relation to atmospheric forcing.

Deep-Sea Res I 47:655–680

Bourke RH, Weigel AM, Paquette RG (1988) The westward turning branch of the West Spitsbergen Current. J Geo- phys Res 93:14065–14077

Brown RGB (1976) The foraging range of breeding dovekies Alle alle.Can Field Nat 90:166–168

Dickson RR, Osborn TJ, Hurrell JW, Meincke J and others (2000) The Arctic Ocean response to the North Atlantic Oscillation. J Clim 13:2671–2696

Egevang C, Boertmann D, Mosbech A, Tamstorf MP (2003) Estimating colony area and population size of little auks Alle alleat Northumberland Island using aerial images.

Polar Biol 26:8–13

Falk K, Pedersen CE, Kampp K (2000) Measurements of diving depth in dovekies (Alle alle). Auk 117:522–525 Falk-Petersen S, Pavlov V, Timofeev S, Sargent JR (2007) Cli-

mate variability and possible effects on Arctic food chains:

the role of Calanus. In: Ørbæk JB, Kallenborn R, Tombre I (eds) Arctic alpine ecosystems and people in a changing environment. Springer-Verlag, Berlin, p 147–166 Gall AE, Roby DD, Irons DB, Rose IC (2006) Differential

response in chick survival to diet in least and crested auklets. Mar Ecol Prog Ser 308:279–291

Golet GL, Kuletz KJ, Roby DD, Irons DB (2000) Adult prey choice affects chick growth and reproductive success in pigeon guillemots. Auk 117:82–91

Hanssen H (1997) Mesozooplankton of the Laptev Sea and the adjacent eastern Nansen Basin — distribution and community structure in late summer. Rep Polar Res 229:

1–131

Harding AMA, Van Pelt TI, Lifjeld JT, Mehlum F (2004) Sex differences in little auk Alle alleparental care: transition from biparental to paternal-only care. Ibis 146:642–651 Harding AMA, Egevang C, Walkusz W, Merkel F, Blanc S,

Grémillet D (2009a) Estimating prey capture rates of a planktivorous seabird, the little auk (Alle alle), using diet, diving behaviour, and energy consumption. Polar Biol 32:

785–796

Harding AMA, Kitaysky AS, Hall ME, Welcker J and others (2009b) Flexibility in the parental effort of an Arctic- breeding seabird. Funct Ecol 23:348–358

Hedd A, Ryder JL, Cowen LL, Bertram DF (2002) Inter-annual variation in the diet, provisioning and growth of Cassin’s auklet at Triangle Island, British Columbia: responses to variation in ocean climate. Mar Ecol Prog Ser 229:221–232 Helaouët P, Beaugrand G (2007) Macroecology of Calanus finmarchicusand C. helgolandicusin the North Atlantic Ocean and adjacent seas. Mar Ecol Prog Ser 345:147–165 Hirche HJ (1997) Life cycle of the copepod Calanus hyper-

boreusin the Greenland Sea. Mar Biol 128:607–618

➤

(11)

Hirche HJ, Kosobokova KN (2007) Distribution of Calanus fin- marchicus in the northern North Atlantic and Arctic Ocean—expatriation and potential colonization. Deep- Sea Res II 54:2729–2747

Hop H, Falk-Petersen S, Svendsen H, Kwas´niewski S, Pavlov V, Pavlova O, Soreide JE (2006) Physical and biological characteristics of the pelagic system across Fram Strait to Kongsfjorden. Prog Oceanogr 71:182–231

Hunt GL Jr, Harrison NM (1990) Foraging habitat and prey taken by least auklets at King Island, Alaska. Mar Ecol Prog Ser 65:141–150

Hurrell JW (1995) Decadal trends in the North Atlantic Oscil- lation: regional temperatures and precipitation. Science 269:676–679

Isaksen K, Gavrilo MV (2000) Little auk Alle alle. In: Anker- Nilssen T, Bakken V, Strøm H, Golovkin AN, Bianki VV, Tatarinkova IP (eds) The status of marine birds breeding in the Barents Sea region. Norsk Polarinstitutt, Tromso, p 131–136

Jakubas D, Wojczulanis K, Walkusz W (2007) Response of dovekie to changes in food availability. Waterbirds 30:

421–428

Kampp K, Meltofte H, Mortensen CE (1987) Population size of the little auk Alle alle in East Greenland. Dan Ornitol Foren Tidssk 81:129–136

Karnovsky NJ, Hunt GL Jr (2002) Estimation of carbon flux to dovekies (Alle alle)in the North Water. Deep-Sea Res II 49:5117–5130

Karnovsky NJ, Kwas´niewski S, We˛s8awski JM, Walkusz W, Beszczy?ska-Möller A (2003) Foraging behavior of the little auks in a heterogeneous environment. Mar Ecol Prog Ser 253:289–303

Karnovsky NJ, Hobson KA, Iverson S, Hunt GL Jr (2008) Sea- sonal changes in the diets of seabirds in the North Water Polynya: a multiple-indicator approach. Mar Ecol Prog Ser 357:291–299

Loeng H, Ozhigin V, Adlandsvik B (1997) Water fluxes through the Barents Sea. ICES J Mar Sci 54:310–317 Lønne OJ, Gulliksen B (1989) Size, age and diet of polar cod,

Boreogadus saida(Lepechin 1773), in ice covered waters.

Polar Biol 9:187–191

Moline MA, Karnovsky NJ, Brown Z, Divoky GJ and others (2008) High latitude changes in ice dynamics and their impact on polar marine ecosystems. Ann N Y Acad Sci 1134:267–319

Mumm N (1991) On the summerly distribution of mesozooplankton in the Nansen Basin, Arctic Ocean. Ber Polar- forsch 92:1–146

Österblom H, Olsson O, Blenckner T, Furness RW (2008) Junk-food in marine ecosystems. Oikos 117:967–977 Pedersen CE, Falk K (2001) Chick diet of dovekies Alle allein

Northwest Greenland. Polar Biol 24:53–58

Rayner NA, Parker DE, Horton EB, Folland CK and others (2003) Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nine- teenth century. J Geophys Res 108:4407 doi:10.1029/2002 JD002670

Roby DD, Brink KL, Nettleship DN (1981) Measurements, chick meals and breeding distribution of dovekies (Alle alle)in Northwest Greenland. Arctic 34:241–248

Roeckner E, Oberhuber JM, Bacher A, Christoph M, Kirchner I (1996) ENSO variability and atmospheric response in a global coupled atmosphere–ocean GCM. Clim Dyn 12:

737–754

Saloranta TM, Svendsen H (2001) Across the Arctic front west of Spitsbergen: high resolution CTD sections from 1998–2000. Polar Res 20:177–184

Schlichtholz P, Goszczko I (2006) Interannual variability of the Atlantic water layer in the West Spitsbergen Current at 76.5° N in summer 1991–2003. Deep-Sea Res I 53:

608–626

Scott CL, Kwas´niewski S, Falk-Petersen S, Sargent JR (2000) Lipids and life strategies of Calanus finmarchicus, Calanus glacialis and Calanus hyperboreus in late autumn, Kongsfjorden, Svalbard. Polar Biol 23:510–516 Stempniewicz L (2001) Alle alle little auk. BWP update.

J Birds West Palearctic 3:175–201

Stempniewicz L, B8achowiak-Samo8yk K, We¸8sawski JM (2007) Impact of climate change on zooplankton commu- nities, seabird populations and arctic terrestrial ecosystem — a scenario. Deep-Sea Res I 54:2934–2945

Swerpel S, Zajaczkowski M (1990) Physical environment of South Spitsbergen. In: Klekowski RZ, Weslawski, JM (eds) Atlas of the marine fauna of Southern Spitsbergen. Osso- lineum, Wroclaw, p 25–41

Tranter DJ, Fraser H (1968) Zooplankton sampling. Mono- graphs on oceanographic methodology 2. UNESCO, Paris Velando A, Alonso-Alvarez C (2003) Differential body condition regulation by males and females in response to exper- imental manipulation of brood size and parental effort in the blue-footed booby. J Anim Ecol 72:846–856

Vinje T (2001) Anomalies and trends of sea-ice extent and atmospheric circulation in the Nordic Seas during the period 1864–1998. J Clim 14:255–267

Visbeck MH, Hurrell JW, Polvani L, Cullen HM (2001) The North Atlantic Oscillation: past, present, and future. Proc Natl Acad Sci USA 98:12876–12877

Walczowski W, Piechura J (2006) New evidence of warming propagating toward the Arctic Ocean. Geophys Res Lett 33:L12601 doi:10.1029/2006GL025872

Wanless S, Harris MP, Redman P, Speakman JR (2005) Low energy value of fish as a probable cause of a major seabird breeding failure in the North Sea. Mar Ecol Prog Ser 294:

1–8

Warner AJ, Hays GC (1994) Sampling by the continuous plankton recorder survey. Prog Oceanogr 34:237–256 Welcker J, Harding AMA, Kitaysky AS, Speakman JR,

Gabrielsen GW (2009a) Daily energy expenditure increases in response to high food availability in an Arctic- breeding seabird with no negative effects on mortality.

Funct Ecol 23:1081–1090

Welcker J, Harding AMA, Karnovsky NJ, Steen H, Strøm H, Gabrielsen GW (2009b) Flexibility in the bimodal foraging strategy of the high Arctic alcid, the little auk, Alle alle.

J Avian Biol 40:388–399

We¸8sawski JM, Stempniewicz L, Galaktionov KM (1994) The food and feeding of little auk from Franz Josef Land. Polar Res 13:173–181

We¸8sawski JM, Koszteyn J, Kwas´niewski S, Stempniewicz L, Malinga M (1999a) Summer food resources of the little auk, Alle alle(L.) in the European Arctic seas. Pol Polar Res 10:387–403

We¸8sawski JM, Stempniewicz L, Mehlum F, Kwas´niewski S (1999b) Summer feeding strategy of the little auk (Alle alle) from Bjornoya, Barents Sea. Polar Biol 21:129–134 Woodruff SD, Slutz RJ, Jenne RL, Steurer PM (1987) A com-

prehensive ocean–atmosphere data set. Bull Am Meteorol Soc 68:1239–1250

Editorial responsibility: Yves Cherel, Villiers-en-Bois, France

Submitted: May 28, 2009; Accepted: July 28, 2010 Proofs received from author(s): September 14, 2010

➤